Why
The current baseline starts with an empty corpus (valid for the guidelines), but optional shared seeds would improve comparability and reproducibility when a seed set is desired.
Scope
- Keep the default empty corpus behavior unchanged.
- Add an optional shared seed corpus input (local path or S3 prefix).
- Copy seeds into each fuzzer's corpus directory before the run.
- Record seed corpus path, size, and hash in the run manifest/report.
Guideline refs